diff options
Diffstat (limited to 'net/sched/sch_htb.c')
| -rw-r--r-- | net/sched/sch_htb.c | 1971 |
1 files changed, 916 insertions, 1055 deletions
diff --git a/net/sched/sch_htb.c b/net/sched/sch_htb.c index 558cc087e60..9f949abcace 100644 --- a/net/sched/sch_htb.c +++ b/net/sched/sch_htb.c @@ -1,4 +1,4 @@ -/* vim: ts=8 sw=8 +/* * net/sched/sch_htb.c Hierarchical token bucket, feed tree version * * This program is free software; you can redistribute it and/or @@ -11,7 +11,7 @@ * Credits (in time order) for older HTB versions: * Stef Coene <stef.coene@docum.org> * HTB support at LARTC mailing list - * Ondrej Kraus, <krauso@barr.cz> + * Ondrej Kraus, <krauso@barr.cz> * found missing INIT_QDISC(htb) * Vladimir Smelhaus, Aamer Akhter, Bert Hubert * helped a lot to locate nasty class stall bug @@ -24,267 +24,169 @@ * Jiri Fojtasek * fixed requeue routine * and many others. thanks. - * - * $Id: sch_htb.c,v 1.25 2003/12/07 11:08:25 devik Exp devik $ */ -#include <linux/config.h> #include <linux/module.h> -#include <asm/uaccess.h> -#include <asm/system.h> -#include <linux/bitops.h> +#include <linux/moduleparam.h> #include <linux/types.h> #include <linux/kernel.h> -#include <linux/sched.h> #include <linux/string.h> -#include <linux/mm.h> -#include <linux/socket.h> -#include <linux/sockios.h> -#include <linux/in.h> #include <linux/errno.h> -#include <linux/interrupt.h> -#include <linux/if_ether.h> -#include <linux/inet.h> -#include <linux/netdevice.h> -#include <linux/etherdevice.h> -#include <linux/notifier.h> -#include <net/ip.h> -#include <net/route.h> #include <linux/skbuff.h> #include <linux/list.h> #include <linux/compiler.h> -#include <net/sock.h> -#include <net/pkt_sched.h> #include <linux/rbtree.h> +#include <linux/workqueue.h> +#include <linux/slab.h> +#include <net/netlink.h> +#include <net/sch_generic.h> +#include <net/pkt_sched.h> /* HTB algorithm. Author: devik@cdi.cz ======================================================================== HTB is like TBF with multiple classes. It is also similar to CBQ because - it allows to assign priority to each class in hierarchy. + it allows to assign priority to each class in hierarchy. In fact it is another implementation of Floyd's formal sharing. Levels: - Each class is assigned level. Leaf has ALWAYS level 0 and root + Each class is assigned level. Leaf has ALWAYS level 0 and root classes have level TC_HTB_MAXDEPTH-1. Interior nodes has level one less than their parent. */ -#define HTB_HSIZE 16 /* classid hash size */ -#define HTB_EWMAC 2 /* rate average over HTB_EWMAC*HTB_HSIZE sec */ -#undef HTB_DEBUG /* compile debugging support (activated by tc tool) */ -#define HTB_RATECM 1 /* whether to use rate computer */ -#define HTB_HYSTERESIS 1/* whether to use mode hysteresis for speedup */ -#define HTB_QLOCK(S) spin_lock_bh(&(S)->dev->queue_lock) -#define HTB_QUNLOCK(S) spin_unlock_bh(&(S)->dev->queue_lock) -#define HTB_VER 0x30011 /* major must be matched with number suplied by TC as version */ +static int htb_hysteresis __read_mostly = 0; /* whether to use mode hysteresis for speedup */ +#define HTB_VER 0x30011 /* major must be matched with number suplied by TC as version */ #if HTB_VER >> 16 != TC_HTB_PROTOVER #error "Mismatched sch_htb.c and pkt_sch.h" #endif -/* debugging support; S is subsystem, these are defined: - 0 - netlink messages - 1 - enqueue - 2 - drop & requeue - 3 - dequeue main - 4 - dequeue one prio DRR part - 5 - dequeue class accounting - 6 - class overlimit status computation - 7 - hint tree - 8 - event queue - 10 - rate estimator - 11 - classifier - 12 - fast dequeue cache - - L is level; 0 = none, 1 = basic info, 2 = detailed, 3 = full - q->debug uint32 contains 16 2-bit fields one for subsystem starting - from LSB - */ -#ifdef HTB_DEBUG -#define HTB_DBG_COND(S,L) (((q->debug>>(2*S))&3) >= L) -#define HTB_DBG(S,L,FMT,ARG...) if (HTB_DBG_COND(S,L)) \ - printk(KERN_DEBUG FMT,##ARG) -#define HTB_CHCL(cl) BUG_TRAP((cl)->magic == HTB_CMAGIC) -#define HTB_PASSQ q, -#define HTB_ARGQ struct htb_sched *q, -#define static -#undef __inline__ -#define __inline__ -#undef inline -#define inline -#define HTB_CMAGIC 0xFEFAFEF1 -#define htb_safe_rb_erase(N,R) do { BUG_TRAP((N)->rb_color != -1); \ - if ((N)->rb_color == -1) break; \ - rb_erase(N,R); \ - (N)->rb_color = -1; } while (0) -#else -#define HTB_DBG_COND(S,L) (0) -#define HTB_DBG(S,L,FMT,ARG...) -#define HTB_PASSQ -#define HTB_ARGQ -#define HTB_CHCL(cl) -#define htb_safe_rb_erase(N,R) rb_erase(N,R) -#endif +/* Module parameter and sysfs export */ +module_param (htb_hysteresis, int, 0640); +MODULE_PARM_DESC(htb_hysteresis, "Hysteresis mode, less CPU load, less accurate"); +static int htb_rate_est = 0; /* htb classes have a default rate estimator */ +module_param(htb_rate_est, int, 0640); +MODULE_PARM_DESC(htb_rate_est, "setup a default rate estimator (4sec 16sec) for htb classes"); /* used internaly to keep status of single class */ enum htb_cmode { - HTB_CANT_SEND, /* class can't send and can't borrow */ - HTB_MAY_BORROW, /* class can't send but may borrow */ - HTB_CAN_SEND /* class can send */ + HTB_CANT_SEND, /* class can't send and can't borrow */ + HTB_MAY_BORROW, /* class can't send but may borrow */ + HTB_CAN_SEND /* class can send */ }; -/* interior & leaf nodes; props specific to leaves are marked L: */ -struct htb_class -{ -#ifdef HTB_DEBUG - unsigned magic; -#endif - /* general class parameters */ - u32 classid; - struct gnet_stats_basic bstats; - struct gnet_stats_queue qstats; - struct gnet_stats_rate_est rate_est; - struct tc_htb_xstats xstats;/* our special stats */ - int refcnt; /* usage count of this class */ - -#ifdef HTB_RATECM - /* rate measurement counters */ - unsigned long rate_bytes,sum_bytes; - unsigned long rate_packets,sum_packets; -#endif +struct htb_prio { + union { + struct rb_root row; + struct rb_root feed; + }; + struct rb_node *ptr; + /* When class changes from state 1->2 and disconnects from + * parent's feed then we lost ptr value and start from the + * first child again. Here we store classid of the + * last valid ptr (used when ptr is NULL). + */ + u32 last_ptr_id; +}; - /* topology */ - int level; /* our level (see above) */ - struct htb_class *parent; /* parent class */ - struct list_head hlist; /* classid hash list item */ - struct list_head sibling; /* sibling list item */ - struct list_head children; /* children list */ - - union { - struct htb_class_leaf { - struct Qdisc *q; - int prio; - int aprio; - int quantum; - int deficit[TC_HTB_MAXDEPTH]; - struct list_head drop_list; - } leaf; - struct htb_class_inner { - struct rb_root feed[TC_HTB_NUMPRIO]; /* feed trees */ - struct rb_node *ptr[TC_HTB_NUMPRIO]; /* current class ptr */ - /* When class changes from state 1->2 and disconnects from - parent's feed then we lost ptr value and start from the - first child again. Here we store classid of the - last valid ptr (used when ptr is NULL). */ - u32 last_ptr_id[TC_HTB_NUMPRIO]; - } inner; - } un; - struct rb_node node[TC_HTB_NUMPRIO]; /* node for self or feed tree */ - struct rb_node pq_node; /* node for event queue */ - unsigned long pq_key; /* the same type as jiffies global */ - - int prio_activity; /* for which prios are we active */ - enum htb_cmode cmode; /* current mode of the class */ - - /* class attached filters */ - struct tcf_proto *filter_list; - int filter_cnt; - - int warned; /* only one warning about non work conserving .. */ - - /* token bucket parameters */ - struct qdisc_rate_table *rate; /* rate table of the class itself */ - struct qdisc_rate_table *ceil; /* ceiling rate (limits borrows too) */ - long buffer,cbuffer; /* token bucket depth/rate */ - long mbuffer; /* max wait time */ - long tokens,ctokens; /* current number of tokens */ - psched_time_t t_c; /* checkpoint time */ +/* interior & leaf nodes; props specific to leaves are marked L: + * To reduce false sharing, place mostly read fields at beginning, + * and mostly written ones at the end. + */ +struct htb_class { + struct Qdisc_class_common common; + struct psched_ratecfg rate; + struct psched_ratecfg ceil; + s64 buffer, cbuffer;/* token bucket depth/rate */ + s64 mbuffer; /* max wait time */ + u32 prio; /* these two are used only by leaves... */ + int quantum; /* but stored for parent-to-leaf return */ + + struct tcf_proto *filter_list; /* class attached filters */ + int filter_cnt; + int refcnt; /* usage count of this class */ + + int level; /* our level (see above) */ + unsigned int children; + struct htb_class *parent; /* parent class */ + + struct gnet_stats_rate_est64 rate_est; + + /* + * Written often fields + */ + struct gnet_stats_basic_packed bstats; + struct gnet_stats_queue qstats; + struct tc_htb_xstats xstats; /* our special stats */ + + /* token bucket parameters */ + s64 tokens, ctokens;/* current number of tokens */ + s64 t_c; /* checkpoint time */ + + union { + struct htb_class_leaf { + struct list_head drop_list; + int deficit[TC_HTB_MAXDEPTH]; + struct Qdisc *q; + } leaf; + struct htb_class_inner { + struct htb_prio clprio[TC_HTB_NUMPRIO]; + } inner; + } un; + s64 pq_key; + + int prio_activity; /* for which prios are we active */ + enum htb_cmode cmode; /* current mode of the class */ + struct rb_node pq_node; /* node for event queue */ + struct rb_node node[TC_HTB_NUMPRIO]; /* node for self or feed tree */ }; -/* TODO: maybe compute rate when size is too large .. or drop ? */ -static __inline__ long L2T(struct htb_class *cl,struct qdisc_rate_table *rate, - int size) -{ - int slot = size >> rate->rate.cell_log; - if (slot > 255) { - cl->xstats.giants++; - slot = 255; - } - return rate->data[slot]; -} +struct htb_level { + struct rb_root wait_pq; + struct htb_prio hprio[TC_HTB_NUMPRIO]; +}; -struct htb_sched -{ - struct list_head root; /* root classes list */ - struct list_head hash[HTB_HSIZE]; /* hashed by classid */ - struct list_head drops[TC_HTB_NUMPRIO]; /* active leaves (for drops) */ - - /* self list - roots of self generating tree */ - struct rb_root row[TC_HTB_MAXDEPTH][TC_HTB_NUMPRIO]; - int row_mask[TC_HTB_MAXDEPTH]; - struct rb_node *ptr[TC_HTB_MAXDEPTH][TC_HTB_NUMPRIO]; - u32 last_ptr_id[TC_HTB_MAXDEPTH][TC_HTB_NUMPRIO]; - - /* self wait list - roots of wait PQs per row */ - struct rb_root wait_pq[TC_HTB_MAXDEPTH]; - - /* time of nearest event per level (row) */ - unsigned long near_ev_cache[TC_HTB_MAXDEPTH]; - - /* cached value of jiffies in dequeue */ - unsigned long jiffies; - - /* whether we hit non-work conserving class during this dequeue; we use */ - int nwc_hit; /* this to disable mindelay complaint in dequeue */ - - int defcls; /* class where unclassified flows go to */ - u32 debug; /* subsystem debug levels */ - - /* filters for qdisc itself */ - struct tcf_proto *filter_list; - int filter_cnt; - - int rate2quantum; /* quant = rate / rate2quantum */ - psched_time_t now; /* cached dequeue time */ - struct timer_list timer; /* send delay timer */ -#ifdef HTB_RATECM - struct timer_list rttim; /* rate computer timer */ - int recmp_bucket; /* which hash bucket to recompute next */ -#endif - - /* non shaped skbs; let them go directly thru */ - struct sk_buff_head direct_queue; - int direct_qlen; /* max qlen of above */ +struct htb_sched { + struct Qdisc_class_hash clhash; + int defcls; /* class where unclassified flows go to */ + int rate2quantum; /* quant = rate / rate2quantum */ - long direct_pkts; -}; + /* filters for qdisc itself */ + struct tcf_proto *filter_list; -/* compute hash of size HTB_HSIZE for given handle */ -static __inline__ int htb_hash(u32 h) -{ -#if HTB_HSIZE != 16 - #error "Declare new hash for your HTB_HSIZE" -#endif - h ^= h>>8; /* stolen from cbq_hash */ - h ^= h>>4; - return h & 0xf; -} +#define HTB_WARN_TOOMANYEVENTS 0x1 + unsigned int warned; /* only one warning */ + int direct_qlen; + struct work_struct work; + + /* non shaped skbs; let them go directly thru */ + struct sk_buff_head direct_queue; + long direct_pkts; + + struct qdisc_watchdog watchdog; + + s64 now; /* cached dequeue time */ + struct list_head drops[TC_HTB_NUMPRIO];/* active leaves (for drops) */ + + /* time of nearest event per level (row) */ + s64 near_ev_cache[TC_HTB_MAXDEPTH]; + + int row_mask[TC_HTB_MAXDEPTH]; + + struct htb_level hlevel[TC_HTB_MAXDEPTH]; +}; /* find class in global hash table using given handle */ -static __inline__ struct htb_class *htb_find(u32 handle, struct Qdisc *sch) +static inline struct htb_class *htb_find(u32 handle, struct Qdisc *sch) { struct htb_sched *q = qdisc_priv(sch); - struct list_head *p; - if (TC_H_MAJ(handle) != sch->handle) + struct Qdisc_class_common *clc; + + clc = qdisc_class_find(&q->clhash, handle); + if (clc == NULL) return NULL; - - list_for_each (p,q->hash+htb_hash(handle)) { - struct htb_class *cl = list_entry(p,struct htb_class,hlist); - if (cl->classid == handle) - return cl; - } - return NULL; + return container_of(clc, struct htb_class, common); } /** @@ -295,17 +197,14 @@ static __inline__ struct htb_class *htb_find(u32 handle, struct Qdisc *sch) * We allow direct class selection by classid in priority. The we examine * filters in qdisc and in inner nodes (if higher filter points to the inner * node). If we end up with classid MAJOR:0 we enqueue the skb into special - * internal fifo (direct). These packets then go directly thru. If we still - * have no valid leaf we try to use MAJOR:default leaf. It still unsuccessfull + * internal fifo (direct). These packets then go directly thru. If we still + * have no valid leaf we try to use MAJOR:default leaf. It still unsuccessful * then finish and return direct queue. */ -#define HTB_DIRECT (struct htb_class*)-1 -static inline u32 htb_classid(struct htb_class *cl) -{ - return (cl && cl != HTB_DIRECT) ? cl->classid : TC_H_UNSPEC; -} +#define HTB_DIRECT ((struct htb_class *)-1L) -static struct htb_class *htb_classify(struct sk_buff *skb, struct Qdisc *sch, int *qerr) +static struct htb_class *htb_classify(struct sk_buff *skb, struct Qdisc *sch, + int *qerr) { struct htb_sched *q = qdisc_priv(sch); struct htb_class *cl; @@ -314,119 +213,72 @@ static struct htb_class *htb_classify(struct sk_buff *skb, struct Qdisc *sch, in int result; /* allow to select class by setting skb->priority to valid classid; - note that nfmark can be used too by attaching filter fw with no - rules in it */ + * note that nfmark can be used too by attaching filter fw with no + * rules in it + */ if (skb->priority == sch->handle) - return HTB_DIRECT; /* X:0 (direct flow) selected */ - if ((cl = htb_find(skb->priority,sch)) != NULL && cl->level == 0) - return cl; + return HTB_DIRECT; /* X:0 (direct flow) selected */ + cl = htb_find(skb->priority, sch); + if (cl) { + if (cl->level == 0) + return cl; + /* Start with inner filter chain if a non-leaf class is selected */ + tcf = cl->filter_list; + } else { + tcf = q->filter_list; + } - *qerr = NET_XMIT_DROP; - tcf = q->filter_list; + *qerr = NET_XMIT_SUCCESS | __NET_XMIT_BYPASS; while (tcf && (result = tc_classify(skb, tcf, &res)) >= 0) { #ifdef CONFIG_NET_CLS_ACT switch (result) { case TC_ACT_QUEUED: - case TC_ACT_STOLEN: - *qerr = NET_XMIT_SUCCESS; + case TC_ACT_STOLEN: + *qerr = NET_XMIT_SUCCESS | __NET_XMIT_STOLEN; case TC_ACT_SHOT: return NULL; } -#elif defined(CONFIG_NET_CLS_POLICE) - if (result == TC_POLICE_SHOT) - return HTB_DIRECT; #endif - if ((cl = (void*)res.class) == NULL) { + cl = (void *)res.class; + if (!cl) { if (res.classid == sch->handle) - return HTB_DIRECT; /* X:0 (direct flow) */ - if ((cl = htb_find(res.classid,sch)) == NULL) - break; /* filter selected invalid classid */ + return HTB_DIRECT; /* X:0 (direct flow) */ + cl = htb_find(res.classid, sch); + if (!cl) + break; /* filter selected invalid classid */ } if (!cl->level) - return cl; /* we hit leaf; return it */ + return cl; /* we hit leaf; return it */ /* we have got inner class; apply inner filter chain */ tcf = cl->filter_list; } /* classification failed; try to use default class */ - cl = htb_find(TC_H_MAKE(TC_H_MAJ(sch->handle),q->defcls),sch); + cl = htb_find(TC_H_MAKE(TC_H_MAJ(sch->handle), q->defcls), sch); if (!cl || cl->level) - return HTB_DIRECT; /* bad default .. this is safe bet */ + return HTB_DIRECT; /* bad default .. this is safe bet */ return cl; } -#ifdef HTB_DEBUG -static void htb_next_rb_node(struct rb_node **n); -#define HTB_DUMTREE(root,memb) if(root) { \ - struct rb_node *n = (root)->rb_node; \ - while (n->rb_left) n = n->rb_left; \ - while (n) { \ - struct htb_class *cl = rb_entry(n, struct htb_class, memb); \ - printk(" %x",cl->classid); htb_next_rb_node (&n); \ - } } - -static void htb_debug_dump (struct htb_sched *q) -{ - int i,p; - printk(KERN_DEBUG "htb*g j=%lu lj=%lu\n",jiffies,q->jiffies); - /* rows */ - for (i=TC_HTB_MAXDEPTH-1;i>=0;i--) { - printk(KERN_DEBUG "htb*r%d m=%x",i,q->row_mask[i]); - for (p=0;p<TC_HTB_NUMPRIO;p++) { - if (!q->row[i][p].rb_node) continue; - printk(" p%d:",p); - HTB_DUMTREE(q->row[i]+p,node[p]); - } - printk("\n"); - } - /* classes */ - for (i = 0; i < HTB_HSIZE; i++) { - struct list_head *l; - list_for_each (l,q->hash+i) { - struct htb_class *cl = list_entry(l,struct htb_class,hlist); - long diff = PSCHED_TDIFF_SAFE(q->now, cl->t_c, (u32)cl->mbuffer); - printk(KERN_DEBUG "htb*c%x m=%d t=%ld c=%ld pq=%lu df=%ld ql=%d " - "pa=%x f:", - cl->classid,cl->cmode,cl->tokens,cl->ctokens, - cl->pq_node.rb_color==-1?0:cl->pq_key,diff, - cl->level?0:cl->un.leaf.q->q.qlen,cl->prio_activity); - if (cl->level) - for (p=0;p<TC_HTB_NUMPRIO;p++) { - if (!cl->un.inner.feed[p].rb_node) continue; - printk(" p%d a=%x:",p,cl->un.inner.ptr[p]?rb_entry(cl->un.inner.ptr[p], struct htb_class,node[p])->classid:0); - HTB_DUMTREE(cl->un.inner.feed+p,node[p]); - } - printk("\n"); - } - } -} -#endif /** * htb_add_to_id_tree - adds class to the round robin list * * Routine adds class to the list (actually tree) sorted by classid. * Make sure that class is not already on such list for given prio. */ -static void htb_add_to_id_tree (HTB_ARGQ struct rb_root *root, - struct htb_class *cl,int prio) +static void htb_add_to_id_tree(struct rb_root *root, + struct htb_class *cl, int prio) { struct rb_node **p = &root->rb_node, *parent = NULL; - HTB_DBG(7,3,"htb_add_id_tree cl=%X prio=%d\n",cl->classid,prio); -#ifdef HTB_DEBUG - if (cl->node[prio].rb_color != -1) { BUG_TRAP(0); return; } - HTB_CHCL(cl); - if (*p) { - struct htb_class *x = rb_entry(*p,struct htb_class,node[prio]); - HTB_CHCL(x); - } -#endif + while (*p) { - struct htb_class *c; parent = *p; + struct htb_class *c; + parent = *p; c = rb_entry(parent, struct htb_class, node[prio]); - HTB_CHCL(c); - if (cl->classid > c->classid) + + if (cl->common.classid > c->common.classid) p = &parent->rb_right; - else + else p = &parent->rb_left; } rb_link_node(&cl->node[prio], parent, p); @@ -440,35 +292,30 @@ static void htb_add_to_id_tree (HTB_ARGQ struct rb_root *root, * change its mode in cl->pq_key microseconds. Make sure that class is not * already in the queue. */ -static void htb_add_to_wait_tree (struct htb_sched *q, - struct htb_class *cl,long delay,int debug_hint) +static void htb_add_to_wait_tree(struct htb_sched *q, + struct htb_class *cl, s64 delay) { - struct rb_node **p = &q->wait_pq[cl->level].rb_node, *parent = NULL; - HTB_DBG(7,3,"htb_add_wt cl=%X key=%lu\n",cl->classid,cl->pq_key); -#ifdef HTB_DEBUG - if (cl->pq_node.rb_color != -1) { BUG_TRAP(0); return; } - HTB_CHCL(cl); - if ((delay <= 0 || delay > cl->mbuffer) && net_ratelimit()) - printk(KERN_ERR "HTB: suspicious delay in wait_tree d=%ld cl=%X h=%d\n",delay,cl->classid,debug_hint); -#endif - cl->pq_key = q->jiffies + PSCHED_US2JIFFIE(delay); - if (cl->pq_key == q->jiffies) + struct rb_node **p = &q->hlevel[cl->level].wait_pq.rb_node, *parent = NULL; + + cl->pq_key = q->now + delay; + if (cl->pq_key == q->now) cl->pq_key++; /* update the nearest event cache */ - if (time_after(q->near_ev_cache[cl->level], cl->pq_key)) + if (q->near_ev_cache[cl->level] > cl->pq_key) q->near_ev_cache[cl->level] = cl->pq_key; - + while (*p) { - struct htb_class *c; parent = *p; + struct htb_class *c; + parent = *p; c = rb_entry(parent, struct htb_class, pq_node); - if (time_after_eq(cl->pq_key, c->pq_key)) + if (cl->pq_key >= c->pq_key) p = &parent->rb_right; - else + else p = &parent->rb_left; } rb_link_node(&cl->pq_node, parent, p); - rb_insert_color(&cl->pq_node, &q->wait_pq[cl->level]); + rb_insert_color(&cl->pq_node, &q->hlevel[cl->level].wait_pq); } /** @@ -477,7 +324,7 @@ static void htb_add_to_wait_tree (struct htb_sched *q, * When we are past last key we return NULL. * Average complexity is 2 steps per call. */ -static void htb_next_rb_node(struct rb_node **n) +static inline void htb_next_rb_node(struct rb_node **n) { *n = rb_next(*n); } @@ -488,42 +335,53 @@ static void htb_next_rb_node(struct rb_node **n) * The class is added to row at priorities marked in mask. * It does nothing if mask == 0. */ -static inline void htb_add_class_to_row(struct htb_sched *q, - struct htb_class *cl,int mask) +static inline void htb_add_class_to_row(struct htb_sched *q, + struct htb_class *cl, int mask) { - HTB_DBG(7,2,"htb_addrow cl=%X mask=%X rmask=%X\n", - cl->classid,mask,q->row_mask[cl->level]); - HTB_CHCL(cl); q->row_mask[cl->level] |= mask; while (mask) { int prio = ffz(~mask); mask &= ~(1 << prio); - htb_add_to_id_tree(HTB_PASSQ q->row[cl->level]+prio,cl,prio); + htb_add_to_id_tree(&q->hlevel[cl->level].hprio[prio].row, cl, prio); } } +/* If this triggers, it is a bug in this code, but it need not be fatal */ +static void htb_safe_rb_erase(struct rb_node *rb, struct rb_root *root) +{ + if (RB_EMPTY_NODE(rb)) { + WARN_ON(1); + } else { + rb_erase(rb, root); + RB_CLEAR_NODE(rb); + } +} + + /** * htb_remove_class_from_row - removes class from its row * * The class is removed from row at priorities marked in mask. * It does nothing if mask == 0. */ -static __inline__ void htb_remove_class_from_row(struct htb_sched *q, - struct htb_class *cl,int mask) +static inline void htb_remove_class_from_row(struct htb_sched *q, + struct htb_class *cl, int mask) { int m = 0; - HTB_CHCL(cl); + struct htb_level *hlevel = &q->hlevel[cl->level]; + while (mask) { int prio = ffz(~mask); + struct htb_prio *hprio = &hlevel->hprio[prio]; + mask &= ~(1 << prio); - if (q->ptr[cl->level][prio] == cl->node+prio) - htb_next_rb_node(q->ptr[cl->level]+prio); - htb_safe_rb_erase(cl->node + prio,q->row[cl->level]+prio); - if (!q->row[cl->level][prio].rb_node) + if (hprio->ptr == cl->node + prio) + htb_next_rb_node(&hprio->ptr); + + htb_safe_rb_erase(cl->node + prio, &hprio->row); + if (!hprio->row.rb_node) m |= 1 << prio; } - HTB_DBG(7,2,"htb_delrow cl=%X mask=%X rmask=%X maskdel=%X\n", - cl->classid,mask,q->row_mask[cl->level],m); q->row_mask[cl->level] &= ~m; } @@ -531,115 +389,123 @@ static __inline__ void htb_remove_class_from_row(struct htb_sched *q, * htb_activate_prios - creates active classe's feed chain * * The class is connected to ancestors and/or appropriate rows - * for priorities it is participating on. cl->cmode must be new + * for priorities it is participating on. cl->cmode must be new * (activated) mode. It does nothing if cl->prio_activity == 0. */ -static void htb_activate_prios(struct htb_sched *q,struct htb_class *cl) +static void htb_activate_prios(struct htb_sched *q, struct htb_class *cl) { struct htb_class *p = cl->parent; - long m,mask = cl->prio_activity; - HTB_DBG(7,2,"htb_act_prios cl=%X mask=%lX cmode=%d\n",cl->classid,mask,cl->cmode); - HTB_CHCL(cl); + long m, mask = cl->prio_activity; while (cl->cmode == HTB_MAY_BORROW && p && mask) { - HTB_CHCL(p); - m = mask; while (m) { + m = mask; + while (m) { int prio = ffz(~m); m &= ~(1 << prio); - - if (p->un.inner.feed[prio].rb_node) + + if (p->un.inner.clprio[prio].feed.rb_node) /* parent already has its feed in use so that - reset bit in mask as parent is already ok */ + * reset bit in mask as parent is already ok + */ mask &= ~(1 << prio); - - htb_add_to_id_tree(HTB_PASSQ p->un.inner.feed+prio,cl,prio); + + htb_add_to_id_tree(&p->un.inner.clprio[prio].feed, cl, prio); } - HTB_DBG(7,3,"htb_act_pr_aft p=%X pact=%X mask=%lX pmode=%d\n", - p->classid,p->prio_activity,mask,p->cmode); p->prio_activity |= mask; - cl = p; p = cl->parent; - HTB_CHCL(cl); + cl = p; + p = cl->parent; + } if (cl->cmode == HTB_CAN_SEND && mask) - htb_add_class_to_row(q,cl,mask); + htb_add_class_to_row(q, cl, mask); } /** * htb_deactivate_prios - remove class from feed chain * - * cl->cmode must represent old mode (before deactivation). It does + * cl->cmode must represent old mode (before deactivation). It does * nothing if cl->prio_activity == 0. Class is removed from all feed * chains and rows. */ static void htb_deactivate_prios(struct htb_sched *q, struct htb_class *cl) { struct htb_class *p = cl->parent; - long m,mask = cl->prio_activity; - HTB_DBG(7,2,"htb_deact_prios cl=%X mask=%lX cmode=%d\n",cl->classid,mask,cl->cmode); - HTB_CHCL(cl); + long m, mask = cl->prio_activity; while (cl->cmode == HTB_MAY_BORROW && p && mask) { - m = mask; mask = 0; + m = mask; + mask = 0; while (m) { int prio = ffz(~m); m &= ~(1 << prio); - - if (p->un.inner.ptr[prio] == cl->node+prio) { + + if (p->un.inner.clprio[prio].ptr == cl->node + prio) { /* we are removing child which is pointed to from - parent feed - forget the pointer but remember - classid */ - p->un.inner.last_ptr_id[prio] = cl->classid; - p->un.inner.ptr[prio] = NULL; + * parent feed - forget the pointer but remember + * classid + */ + p->un.inner.clprio[prio].last_ptr_id = cl->common.classid; + p->un.inner.clprio[prio].ptr = NULL; } - - htb_safe_rb_erase(cl->node + prio,p->un.inner.feed + prio); - - if (!p->un.inner.feed[prio].rb_node) + + htb_safe_rb_erase(cl->node + prio, + &p->un.inner.clprio[prio].feed); + + if (!p->un.inner.clprio[prio].feed.rb_node) mask |= 1 << prio; } - HTB_DBG(7,3,"htb_deact_pr_aft p=%X pact=%X mask=%lX pmode=%d\n", - p->classid,p->prio_activity,mask,p->cmode); + p->prio_activity &= ~mask; - cl = p; p = cl->parent; - HTB_CHCL(cl); + cl = p; + p = cl->parent; + } - if (cl->cmode == HTB_CAN_SEND && mask) - htb_remove_class_from_row(q,cl,mask); + if (cl->cmode == HTB_CAN_SEND && mask) + htb_remove_class_from_row(q, cl, mask); +} + +static inline s64 htb_lowater(const struct htb_class *cl) +{ + if (htb_hysteresis) + return cl->cmode != HTB_CANT_SEND ? -cl->cbuffer : 0; + else + return 0; } +static inline s64 htb_hiwater(const struct htb_class *cl) +{ + if (htb_hysteresis) + return cl->cmode == HTB_CAN_SEND ? -cl->buffer : 0; + else + return 0; +} + /** * htb_class_mode - computes and returns current class mode * * It computes cl's mode at time cl->t_c+diff and returns it. If mode * is not HTB_CAN_SEND then cl->pq_key is updated to time difference - * from now to time when cl will change its state. + * from now to time when cl will change its state. * Also it is worth to note that class mode doesn't change simply - * at cl->{c,}tokens == 0 but there can rather be hysteresis of + * at cl->{c,}tokens == 0 but there can rather be hysteresis of * 0 .. -cl->{c,}buffer range. It is meant to limit number of * mode transitions per time unit. The speed gain is about 1/6. */ -static __inline__ enum htb_cmode -htb_class_mode(struct htb_class *cl,long *diff) +static inline enum htb_cmode +htb_class_mode(struct htb_class *cl, s64 *diff) { - long toks; + s64 toks; - if ((toks = (cl->ctokens + *diff)) < ( -#if HTB_HYSTERESIS - cl->cmode != HTB_CANT_SEND ? -cl->cbuffer : -#endif - 0)) { - *diff = -toks; - return HTB_CANT_SEND; - } - if ((toks = (cl->tokens + *diff)) >= ( -#if HTB_HYSTERESIS - cl->cmode == HTB_CAN_SEND ? -cl->buffer : -#endif - 0)) - return HTB_CAN_SEND; + if ((toks = (cl->ctokens + *diff)) < htb_lowater(cl)) { + *diff = -toks; + return HTB_CANT_SEND; + } - *diff = -toks; - return HTB_MAY_BORROW; + if ((toks = (cl->tokens + *diff)) >= htb_hiwater(cl)) + return HTB_CAN_SEND; + + *diff = -toks; + return HTB_MAY_BORROW; } /** @@ -651,167 +517,118 @@ htb_class_mode(struct htb_class *cl,long *diff) * be different from old one and cl->pq_key has to be valid if changing * to mode other than HTB_CAN_SEND (see htb_add_to_wait_tree). */ -static void -htb_change_class_mode(struct htb_sched *q, struct htb_class *cl, long *diff) -{ - enum htb_cmode new_mode = htb_class_mode(cl,diff); - - HTB_CHCL(cl); - HTB_DBG(7,1,"htb_chging_clmode %d->%d cl=%X\n",cl->cmode,new_mode,cl->classid); +static void +htb_change_class_mode(struct htb_sched *q, struct htb_class *cl, s64 *diff) +{ + enum htb_cmode new_mode = htb_class_mode(cl, diff); if (new_mode == cl->cmode) - return; - - if (cl->prio_activity) { /* not necessary: speed optimization */ - if (cl->cmode != HTB_CANT_SEND) - htb_deactivate_prios(q,cl); + return; + + if (cl->prio_activity) { /* not necessary: speed optimization */ + if (cl->cmode != HTB_CANT_SEND) + htb_deactivate_prios(q, cl); cl->cmode = new_mode; - if (new_mode != HTB_CANT_SEND) - htb_activate_prios(q,cl); - } else + if (new_mode != HTB_CANT_SEND) + htb_activate_prios(q, cl); + } else cl->cmode = new_mode; } /** - * htb_activate - inserts leaf cl into appropriate active feeds + * htb_activate - inserts leaf cl into appropriate active feeds * * Routine learns (new) priority of leaf and activates feed chain * for the prio. It can be called on already active leaf safely. * It also adds leaf into droplist. */ -static __inline__ void htb_activate(struct htb_sched *q,struct htb_class *cl) +static inline void htb_activate(struct htb_sched *q, struct htb_class *cl) { - BUG_TRAP(!cl->level && cl->un.leaf.q && cl->un.leaf.q->q.qlen); - HTB_CHCL(cl); + WARN_ON(cl->level || !cl->un.leaf.q || !cl->un.leaf.q->q.qlen); + if (!cl->prio_activity) { - cl->prio_activity = 1 << (cl->un.leaf.aprio = cl->un.leaf.prio); - htb_activate_prios(q,cl); - list_add_tail(&cl->un.leaf.drop_list,q->drops+cl->un.leaf.aprio); + cl->prio_activity = 1 << cl->prio; + htb_activate_prios(q, cl); + list_add_tail(&cl->un.leaf.drop_list, + q->drops + cl->prio); } } /** - * htb_deactivate - remove leaf cl from active feeds + * htb_deactivate - remove leaf cl from active feeds * * Make sure that leaf is active. In the other words it can't be called * with non-active leaf. It also removes class from the drop list. */ -static __inline__ void -htb_deactivate(struct htb_sched *q,struct htb_class *cl) +static inline void htb_deactivate(struct htb_sched *q, struct htb_class *cl) { - BUG_TRAP(cl->prio_activity); - HTB_CHCL(cl); - htb_deactivate_prios(q,cl); + WARN_ON(!cl->prio_activity); + + htb_deactivate_prios(q, cl); cl->prio_activity = 0; list_del_init(&cl->un.leaf.drop_list); } static int htb_enqueue(struct sk_buff *skb, struct Qdisc *sch) { - int ret; - struct htb_sched *q = qdisc_priv(sch); - struct htb_class *cl = htb_classify(skb,sch,&ret); - - if (cl == HTB_DIRECT) { - /* enqueue to helper queue */ - if (q->direct_queue.qlen < q->direct_qlen) { - __skb_queue_tail(&q->direct_queue, skb); - q->direct_pkts++; - } else { - kfree_skb(skb); - sch->qstats.drops++; - return NET_XMIT_DROP; - } + int uninitialized_var(ret); + struct htb_sched *q = qdisc_priv(sch); + struct htb_class *cl = htb_classify(skb, sch, &ret); + + if (cl == HTB_DIRECT) { + /* enqueue to helper queue */ + if (q->direct_queue.qlen < q->direct_qlen) { + __skb_queue_tail(&q->direct_queue, skb); + q->direct_pkts++; + } else { + return qdisc_drop(skb, sch); + } #ifdef CONFIG_NET_CLS_ACT - } else if (!cl) { - if (ret == NET_XMIT_DROP) - sch->qstats.drops++; - kfree_skb (skb); - return ret; + } else if (!cl) { + if (ret & __NET_XMIT_BYPASS) + sch->qstats.drops++; + kfree_skb(skb); + return ret; #endif - } else if (cl->un.leaf.q->enqueue(skb, cl->un.leaf.q) != NET_XMIT_SUCCESS) { - sch->qstats.drops++; - cl->qstats.drops++; - return NET_XMIT_DROP; - } else { - cl->bstats.packets++; cl->bstats.bytes += skb->len; - htb_activate (q,cl); - } - - sch->q.qlen++; - sch->bstats.packets++; sch->bstats.bytes += skb->len; - HTB_DBG(1,1,"htb_enq_ok cl=%X skb=%p\n",(cl && cl != HTB_DIRECT)?cl->classid:0,skb); - return NET_XMIT_SUCCESS; -} - -/* TODO: requeuing packet charges it to policers again !! */ -static int htb_requeue(struct sk_buff *skb, struct Qdisc *sch) -{ - struct htb_sched *q = qdisc_priv(sch); - int ret = NET_XMIT_SUCCESS; - struct htb_class *cl = htb_classify(skb,sch, &ret); - struct sk_buff *tskb; - - if (cl == HTB_DIRECT || !cl) { - /* enqueue to helper queue */ - if (q->direct_queue.qlen < q->direct_qlen && cl) { - __skb_queue_head(&q->direct_queue, skb); + } else if ((ret = qdisc_enqueue(skb, cl->un.leaf.q)) != NET_XMIT_SUCCESS) { + if (net_xmit_drop_count(ret)) { + sch->qstats.drops++; + cl->qstats.drops++; + } + return ret; } else { - __skb_queue_head(&q->direct_queue, skb); - tskb = __skb_dequeue_tail(&q->direct_queue); - kfree_skb (tskb); - sch->qstats.drops++; - return NET_XMIT_CN; + htb_activate(q, cl); } - } else if (cl->un.leaf.q->ops->requeue(skb, cl->un.leaf.q) != NET_XMIT_SUCCESS) { - sch->qstats.drops++; - cl->qstats.drops++; - return NET_XMIT_DROP; - } else - htb_activate (q,cl); - - sch->q.qlen++; - sch->qstats.requeues++; - HTB_DBG(1,1,"htb_req_ok cl=%X skb=%p\n",(cl && cl != HTB_DIRECT)?cl->classid:0,skb); - return NET_XMIT_SUCCESS; + + sch->q.qlen++; + return NET_XMIT_SUCCESS; } -static void htb_timer(unsigned long arg) +static inline void htb_accnt_tokens(struct htb_class *cl, int bytes, s64 diff) { - struct Qdisc *sch = (struct Qdisc*)arg; - sch->flags &= ~TCQ_F_THROTTLED; - wmb(); - netif_schedule(sch->dev); + s64 toks = diff + cl->tokens; + + if (toks > cl->buffer) + toks = cl->buffer; + toks -= (s64) psched_l2t_ns(&cl->rate, bytes); + if (toks <= -cl->mbuffer) + toks = 1 - cl->mbuffer; + + cl->tokens = toks; } -#ifdef HTB_RATECM -#define RT_GEN(D,R) R+=D-(R/HTB_EWMAC);D=0 -static void htb_rate_timer(unsigned long arg) +static inline void htb_accnt_ctokens(struct htb_class *cl, int bytes, s64 diff) { - struct Qdisc *sch = (struct Qdisc*)arg; - struct htb_sched *q = qdisc_priv(sch); - struct list_head *p; - - /* lock queue so that we can muck with it */ - HTB_QLOCK(sch); - HTB_DBG(10,1,"htb_rttmr j=%ld\n",jiffies); - - q->rttim.expires = jiffies + HZ; - add_timer(&q->rttim); - - /* scan and recompute one bucket at time */ - if (++q->recmp_bucket >= HTB_HSIZE) - q->recmp_bucket = 0; - list_for_each (p,q->hash+q->recmp_bucket) { - struct htb_class *cl = list_entry(p,struct htb_class,hlist); - HTB_DBG(10,2,"htb_rttmr_cl cl=%X sbyte=%lu spkt=%lu\n", - cl->classid,cl->sum_bytes,cl->sum_packets); - RT_GEN (cl->sum_bytes,cl->rate_bytes); - RT_GEN (cl->sum_packets,cl->rate_packets); - } - HTB_QUNLOCK(sch); + s64 toks = diff + cl->ctokens; + + if (toks > cl->cbuffer) + toks = cl->cbuffer; + toks -= (s64) psched_l2t_ns(&cl->ceil, bytes); + if (toks <= -cl->mbuffer) + toks = 1 - cl->mbuffer; + + cl->ctokens = toks; } -#endif /** * htb_charge_class - charges amount "bytes" to leaf and ancestors @@ -824,68 +641,40 @@ static void htb_rate_timer(unsigned long arg) * CAN_SEND) because we can use more precise clock that event queue here. * In such case we remove class from event queue first. */ -static void htb_charge_class(struct htb_sched *q,struct htb_class *cl, - int level,int bytes) -{ - long toks,diff; +static void htb_charge_class(struct htb_sched *q, struct htb_class *cl, + int level, struct sk_buff *skb) +{ + int bytes = qdisc_pkt_len(skb); enum htb_cmode old_mode; - HTB_DBG(5,1,"htb_chrg_cl cl=%X lev=%d len=%d\n",cl->classid,level,bytes); - -#define HTB_ACCNT(T,B,R) toks = diff + cl->T; \ - if (toks > cl->B) toks = cl->B; \ - toks -= L2T(cl, cl->R, bytes); \ - if (toks <= -cl->mbuffer) toks = 1-cl->mbuffer; \ - cl->T = toks + s64 diff; while (cl) { - HTB_CHCL(cl); - diff = PSCHED_TDIFF_SAFE(q->now, cl->t_c, (u32)cl->mbuffer); -#ifdef HTB_DEBUG - if (diff > cl->mbuffer || diff < 0 || PSCHED_TLESS(q->now, cl->t_c)) { - if (net_ratelimit()) - printk(KERN_ERR "HTB: bad diff in charge, cl=%X diff=%lX now=%Lu then=%Lu j=%lu\n", - cl->classid, diff, -#ifdef CONFIG_NET_SCH_CLK_GETTIMEOFDAY - q->now.tv_sec * 1000000ULL + q->now.tv_usec, - cl->t_c.tv_sec * 1000000ULL + cl->t_c.tv_usec, -#else - (unsigned long long) q->now, - (unsigned long long) cl->t_c, -#endif - q->jiffies); - diff = 1000; - } -#endif + diff = min_t(s64, q->now - cl->t_c, cl->mbuffer); if (cl->level >= level) { - if (cl->level == level) cl->xstats.lends++; - HTB_ACCNT (tokens,buffer,rate); + if (cl->level == level) + cl->xstats.lends++; + htb_accnt_tokens(cl, bytes, diff); } else { cl->xstats.borrows++; - cl->tokens += diff; /* we moved t_c; update tokens */ + cl->tokens += diff; /* we moved t_c; update tokens */ } - HTB_ACCNT (ctokens,cbuffer,ceil); + htb_accnt_ctokens(cl, bytes, diff); cl->t_c = q->now; - HTB_DBG(5,2,"htb_chrg_clp cl=%X diff=%ld tok=%ld ctok=%ld\n",cl->classid,diff,cl->tokens,cl->ctokens); - old_mode = cl->cmode; diff = 0; - htb_change_class_mode(q,cl,&diff); + old_mode = cl->cmode; + diff = 0; + htb_change_class_mode(q, cl, &diff); if (old_mode != cl->cmode) { if (old_mode != HTB_CAN_SEND) - htb_safe_rb_erase(&cl->pq_node,q->wait_pq+cl->level); + htb_safe_rb_erase(&cl->pq_node, &q->hlevel[cl->level].wait_pq); if (cl->cmode != HTB_CAN_SEND) - htb_add_to_wait_tree (q,cl,diff,1); + htb_add_to_wait_tree(q, cl, diff); } - -#ifdef HTB_RATECM - /* update rate counters */ - cl->sum_bytes += bytes; cl->sum_packets++; -#endif - /* update byte stats except for leaves which are already updated */ - if (cl->level) { - cl->bstats.bytes += bytes; - cl->bstats.packets++; - } + /* update basic stats except for leaves which are already updated */ + if (cl->level) + bstats_update(&cl->bstats, skb); + cl = cl->parent; } } @@ -893,69 +682,66 @@ static void htb_charge_class(struct htb_sched *q,struct htb_class *cl, /** * htb_do_events - make mode changes to classes at the level * - * Scans event queue for pending events and applies them. Returns jiffies to - * next pending event (0 for no event in pq). - * Note: Aplied are events whose have cl->pq_key <= jiffies. + * Scans event queue for pending events and applies them. Returns time of + * next pending event (0 for no event in pq, q->now for too many events). + * Note: Applied are events whose have cl->pq_key <= q->now. */ -static long htb_do_events(struct htb_sched *q,int level) +static s64 htb_do_events(struct htb_sched *q, const int level, + unsigned long start) { - int i; - HTB_DBG(8,1,"htb_do_events l=%d root=%p rmask=%X\n", - level,q->wait_pq[level].rb_node,q->row_mask[level]); - for (i = 0; i < 500; i++) { + /* don't run for longer than 2 jiffies; 2 is used instead of + * 1 to simplify things when jiffy is going to be incremented + * too soon + */ + unsigned long stop_at = start + 2; + struct rb_root *wait_pq = &q->hlevel[level].wait_pq; + + while (time_before(jiffies, stop_at)) { struct htb_class *cl; - long diff; - struct rb_node *p = q->wait_pq[level].rb_node; - if (!p) return 0; - while (p->rb_left) p = p->rb_left; + s64 diff; + struct rb_node *p = rb_first(wait_pq); + + if (!p) + return 0; cl = rb_entry(p, struct htb_class, pq_node); - if (time_after(cl->pq_key, q->jiffies)) { - HTB_DBG(8,3,"htb_do_ev_ret delay=%ld\n",cl->pq_key - q->jiffies); - return cl->pq_key - q->jiffies; - } - htb_safe_rb_erase(p,q->wait_pq+level); - diff = PSCHED_TDIFF_SAFE(q->now, cl->t_c, (u32)cl->mbuffer); -#ifdef HTB_DEBUG - if (diff > cl->mbuffer || diff < 0 || PSCHED_TLESS(q->now, cl->t_c)) { - if (net_ratelimit()) - printk(KERN_ERR "HTB: bad diff in events, cl=%X diff=%lX now=%Lu then=%Lu j=%lu\n", - cl->classid, diff, -#ifdef CONFIG_NET_SCH_CLK_GETTIMEOFDAY - q->now.tv_sec * 1000000ULL + q->now.tv_usec, - cl->t_c.tv_sec * 1000000ULL + cl->t_c.tv_usec, -#else - (unsigned long long) q->now, - (unsigned long long) cl->t_c, -#endif - q->jiffies); - diff = 1000; - } -#endif - htb_change_class_mode(q,cl,&diff); + if (cl->pq_key > q->now) + return cl->pq_key; + + htb_safe_rb_erase(p, wait_pq); + diff = min_t(s64, q->now - cl->t_c, cl->mbuffer); + htb_change_class_mode(q, cl, &diff); if (cl->cmode != HTB_CAN_SEND) - htb_add_to_wait_tree (q,cl,diff,2); + htb_add_to_wait_tree(q, cl, diff); } - if (net_ratelimit()) - printk(KERN_WARNING "htb: too many events !\n"); - return HZ/10; + + /* too much load - let's continue after a break for scheduling */ + if (!(q->warned & HTB_WARN_TOOMANYEVENTS)) { + pr_warn("htb: too many events!\n"); + q->warned |= HTB_WARN_TOOMANYEVENTS; + } + + return q->now; } /* Returns class->node+prio from id-tree where classe's id is >= id. NULL - is no such one exists. */ -static struct rb_node * -htb_id_find_next_upper(int prio,struct rb_node *n,u32 id) + * is no such one exists. + */ +static struct rb_node *htb_id_find_next_upper(int prio, struct rb_node *n, + u32 id) { struct rb_node *r = NULL; while (n) { - struct htb_class *cl = rb_entry(n,struct htb_class,node[prio]); - if (id == cl->classid) return n; - - if (id > cl->classid) { + struct htb_class *cl = + rb_entry(n, struct htb_class, node[prio]); + + if (id > cl->common.classid) { n = n->rb_right; - } else { + } else if (id < cl->common.classid) { r = n; n = n->rb_left; + } else { + return n; } } return r; @@ -966,226 +752,212 @@ htb_id_find_next_upper(int prio,struct rb_node *n,u32 id) * * Find leaf where current feed pointers points to. */ -static struct htb_class * -htb_lookup_leaf(HTB_ARGQ struct rb_root *tree,int prio,struct rb_node **pptr,u32 *pid) +static struct htb_class *htb_lookup_leaf(struct htb_prio *hprio, const int prio) { int i; struct { struct rb_node *root; struct rb_node **pptr; u32 *pid; - } stk[TC_HTB_MAXDEPTH],*sp = stk; - - BUG_TRAP(tree->rb_node); - sp->root = tree->rb_node; - sp->pptr = pptr; - sp->pid = pid; + } stk[TC_HTB_MAXDEPTH], *sp = stk; + + BUG_ON(!hprio->row.rb_node); + sp->root = hprio->row.rb_node; + sp->pptr = &hprio->ptr; + sp->pid = &hprio->last_ptr_id; for (i = 0; i < 65535; i++) { - HTB_DBG(4,2,"htb_lleaf ptr=%p pid=%X\n",*sp->pptr,*sp->pid); - - if (!*sp->pptr && *sp->pid) { - /* ptr was invalidated but id is valid - try to recover - the original or next ptr */ - *sp->pptr = htb_id_find_next_upper(prio,sp->root,*sp->pid); + if (!*sp->pptr && *sp->pid) { + /* ptr was invalidated but id is valid - try to recover + * the original or next ptr + */ + *sp->pptr = + htb_id_find_next_upper(prio, sp->root, *sp->pid); } - *sp->pid = 0; /* ptr is valid now so that remove this hint as it - can become out of date quickly */ - if (!*sp->pptr) { /* we are at right end; rewind & go up */ + *sp->pid = 0; /* ptr is valid now so that remove this hint as it + * can become out of date quickly + */ + if (!*sp->pptr) { /* we are at right end; rewind & go up */ *sp->pptr = sp->root; - while ((*sp->pptr)->rb_left) + while ((*sp->pptr)->rb_left) *sp->pptr = (*sp->pptr)->rb_left; if (sp > stk) { sp--; - BUG_TRAP(*sp->pptr); if(!*sp->pptr) return NULL; - htb_next_rb_node (sp->pptr); + if (!*sp->pptr) { + WARN_ON(1); + return NULL; + } + htb_next_rb_node(sp->pptr); } } else { struct htb_class *cl; - cl = rb_entry(*sp->pptr,struct htb_class,node[prio]); - HTB_CHCL(cl); - if (!cl->level) + struct htb_prio *clp; + + cl = rb_entry(*sp->pptr, struct htb_class, node[prio]); + if (!cl->level) return cl; - (++sp)->root = cl->un.inner.feed[prio].rb_node; - sp->pptr = cl->un.inner.ptr+prio; - sp->pid = cl->un.inner.last_ptr_id+prio; + clp = &cl->un.inner.clprio[prio]; + (++sp)->root = clp->feed.rb_node; + sp->pptr = &clp->ptr; + sp->pid = &clp->last_ptr_id; } } - BUG_TRAP(0); + WARN_ON(1); return NULL; } /* dequeues packet at given priority and level; call only if - you are sure that there is active class at prio/level */ -static struct sk_buff * -htb_dequeue_tree(struct htb_sched *q,int prio,int level) + * you are sure that there is active class at prio/level + */ +static struct sk_buff *htb_dequeue_tree(struct htb_sched *q, const int prio, + const int level) { struct sk_buff *skb = NULL; - struct htb_class *cl,*start; + struct htb_class *cl, *start; + struct htb_level *hlevel = &q->hlevel[level]; + struct htb_prio *hprio = &hlevel->hprio[prio]; + /* look initial class up in the row */ - start = cl = htb_lookup_leaf (HTB_PASSQ q->row[level]+prio,prio, - q->ptr[level]+prio,q->last_ptr_id[level]+prio); - + start = cl = htb_lookup_leaf(hprio, prio); + do { next: - BUG_TRAP(cl); - if (!cl) return NULL; - HTB_DBG(4,1,"htb_deq_tr prio=%d lev=%d cl=%X defic=%d\n", - prio,level,cl->classid,cl->un.leaf.deficit[level]); + if (unlikely(!cl)) + return NULL; /* class can be empty - it is unlikely but can be true if leaf - qdisc drops packets in enqueue routine or if someone used - graft operation on the leaf since last dequeue; - simply deactivate and skip such class */ + * qdisc drops packets in enqueue routine or if someone used + * graft operation on the leaf since last dequeue; + * simply deactivate and skip such class + */ if (unlikely(cl->un.leaf.q->q.qlen == 0)) { struct htb_class *next; - htb_deactivate(q,cl); + htb_deactivate(q, cl); /* row/level might become empty */ if ((q->row_mask[level] & (1 << prio)) == 0) - return NULL; - - next = htb_lookup_leaf (HTB_PASSQ q->row[level]+prio, - prio,q->ptr[level]+prio,q->last_ptr_id[level]+prio); + return NULL; + + next = htb_lookup_leaf(hprio, prio); - if (cl == start) /* fix start if we just deleted it */ + if (cl == start) /* fix start if we just deleted it */ start = next; cl = next; goto next; } - - if (likely((skb = cl->un.leaf.q->dequeue(cl->un.leaf.q)) != NULL)) + + skb = cl->un.leaf.q->dequeue(cl->un.leaf.q); + if (likely(skb != NULL)) break; - if (!cl->warned) { - printk(KERN_WARNING "htb: class %X isn't work conserving ?!\n",cl->classid); - cl->warned = 1; - } - q->nwc_hit++; - htb_next_rb_node((level?cl->parent->un.inner.ptr:q->ptr[0])+prio); - cl = htb_lookup_leaf (HTB_PASSQ q->row[level]+prio,prio,q->ptr[level]+prio, - q->last_ptr_id[level]+prio); + + qdisc_warn_nonwc("htb", cl->un.leaf.q); + htb_next_rb_node(level ? &cl->parent->un.inner.clprio[prio].ptr: + &q->hlevel[0].hprio[prio].ptr); + cl = htb_lookup_leaf(hprio, prio); } while (cl != start); if (likely(skb != NULL)) { - if ((cl->un.leaf.deficit[level] -= skb->len) < 0) { - HTB_DBG(4,2,"htb_next_cl oldptr=%p quant_add=%d\n", - level?cl->parent->un.inner.ptr[prio]:q->ptr[0][prio],cl->un.leaf.quantum); - cl->un.leaf.deficit[level] += cl->un.leaf.quantum; - htb_next_rb_node((level?cl->parent->un.inner.ptr:q->ptr[0])+prio); + bstats_update(&cl->bstats, skb); + cl->un.leaf.deficit[level] -= qdisc_pkt_len(skb); + if (cl->un.leaf.deficit[level] < 0) { + cl->un.leaf.deficit[level] += cl->quantum; + htb_next_rb_node(level ? &cl->parent->un.inner.clprio[prio].ptr : + &q->hlevel[0].hprio[prio].ptr); } /* this used to be after charge_class but this constelation - gives us slightly better performance */ + * gives us slightly better performance + */ if (!cl->un.leaf.q->q.qlen) - htb_deactivate (q,cl); - htb_charge_class (q,cl,level,skb->len); + htb_deactivate(q, cl); + htb_charge_class(q, cl, level, skb); } return skb; } -static void htb_delay_by(struct Qdisc *sch,long delay) -{ - struct htb_sched *q = qdisc_priv(sch); - if (delay <= 0) delay = 1; - if (unlikely(delay > 5*HZ)) { - if (net_ratelimit()) - printk(KERN_INFO "HTB delay %ld > 5sec\n", delay); - delay = 5*HZ; - } - /* why don't use jiffies here ? because expires can be in past */ - mod_timer(&q->timer, q->jiffies + delay); - sch->flags |= TCQ_F_THROTTLED; - sch->qstats.overlimits++; - HTB_DBG(3,1,"htb_deq t_delay=%ld\n",delay); -} - static struct sk_buff *htb_dequeue(struct Qdisc *sch) { - struct sk_buff *skb = NULL; + struct sk_buff *skb; struct htb_sched *q = qdisc_priv(sch); int level; - long min_delay; -#ifdef HTB_DEBUG - int evs_used = 0; -#endif - - q->jiffies = jiffies; - HTB_DBG(3,1,"htb_deq dircnt=%d qlen=%d\n",skb_queue_len(&q->direct_queue), - sch->q.qlen); + s64 next_event; + unsigned long start_at; /* try to dequeue direct packets as high prio (!) to minimize cpu work */ - if ((skb = __skb_dequeue(&q->direct_queue)) != NULL) { - sch->flags &= ~TCQ_F_THROTTLED; + skb = __skb_dequeue(&q->direct_queue); + if (skb != NULL) { +ok: + qdisc_bstats_update(sch, skb); + qdisc_unthrottled(sch); sch->q.qlen--; return skb; } - if (!sch->q.qlen) goto fin; - PSCHED_GET_TIME(q->now); + if (!sch->q.qlen) + goto fin; + q->now = ktime_to_ns(ktime_get()); + start_at = jiffies; + + next_event = q->now + 5LLU * NSEC_PER_SEC; - min_delay = LONG_MAX; - q->nwc_hit = 0; for (level = 0; level < TC_HTB_MAXDEPTH; level++) { /* common case optimization - skip event handler quickly */ int m; - long delay; - if (time_after_eq(q->jiffies, q->near_ev_cache[level])) { - delay = htb_do_events(q,level); - q->near_ev_cache[level] = q->jiffies + (delay ? delay : HZ); -#ifdef HTB_DEBUG - evs_used++; -#endif - } else - delay = q->near_ev_cache[level] - q->jiffies; - - if (delay && min_delay > delay) - min_delay = delay; + s64 event = q->near_ev_cache[level]; + + if (q->now >= event) { + event = htb_do_events(q, level, start_at); + if (!event) + event = q->now + NSEC_PER_SEC; + q->near_ev_cache[level] = event; + } + + if (next_event > event) + next_event = event; + m = ~q->row_mask[level]; while (m != (int)(-1)) { - int prio = ffz (m); + int prio = ffz(m); + m |= 1 << prio; - skb = htb_dequeue_tree(q,prio,level); - if (likely(skb != NULL)) { - sch->q.qlen--; - sch->flags &= ~TCQ_F_THROTTLED; - goto fin; - } + skb = htb_dequeue_tree(q, prio, level); + if (likely(skb != NULL)) + goto ok; } } -#ifdef HTB_DEBUG - if (!q->nwc_hit && min_delay >= 10*HZ && net_ratelimit()) { - if (min_delay == LONG_MAX) { - printk(KERN_ERR "HTB: dequeue bug (%d,%lu,%lu), report it please !\n", - evs_used,q->jiffies,jiffies); - htb_debug_dump(q); - } else - printk(KERN_WARNING "HTB: mindelay=%ld, some class has " - "too small rate\n",min_delay); + sch->qstats.overlimits++; + if (likely(next_event > q->now)) { + if (!test_bit(__QDISC_STATE_DEACTIVATED, + &qdisc_root_sleeping(q->watchdog.qdisc)->state)) { + ktime_t time = ns_to_ktime(next_event); + qdisc_throttled(q->watchdog.qdisc); + hrtimer_start(&q->watchdog.timer, time, + HRTIMER_MODE_ABS); + } + } else { + schedule_work(&q->work); } -#endif - htb_delay_by (sch,min_delay > 5*HZ ? 5*HZ : min_delay); fin: - HTB_DBG(3,1,"htb_deq_end %s j=%lu skb=%p\n",sch->dev->name,q->jiffies,skb); return skb; } /* try to drop from each class (by prio) until one succeed */ -static unsigned int htb_drop(struct Qdisc* sch) +static unsigned int htb_drop(struct Qdisc *sch) { struct htb_sched *q = qdisc_priv(sch); int prio; for (prio = TC_HTB_NUMPRIO - 1; prio >= 0; prio--) { struct list_head *p; - list_for_each (p,q->drops+prio) { + list_for_each(p, q->drops + prio) { struct htb_class *cl = list_entry(p, struct htb_class, un.leaf.drop_list); unsigned int len; - if (cl->un.leaf.q->ops->drop && - (len = cl->un.leaf.q->ops->drop(cl->un.leaf.q))) { + if (cl->un.leaf.q->ops->drop && + (len = cl->un.leaf.q->ops->drop(cl->un.leaf.q))) { sch->q.qlen--; if (!cl->un.leaf.q->q.qlen) - htb_deactivate (q,cl); + htb_deactivate(q, cl); return len; } } @@ -1195,91 +967,92 @@ static unsigned int htb_drop(struct Qdisc* sch) /* reset all classes */ /* always caled under BH & queue lock */ -static void htb_reset(struct Qdisc* sch) +static void htb_reset(struct Qdisc *sch) { struct htb_sched *q = qdisc_priv(sch); - int i; - HTB_DBG(0,1,"htb_reset sch=%p, handle=%X\n",sch,sch->handle); + struct htb_class *cl; + unsigned int i; - for (i = 0; i < HTB_HSIZE; i++) { - struct list_head *p; - list_for_each (p,q->hash+i) { - struct htb_class *cl = list_entry(p,struct htb_class,hlist); + for (i = 0; i < q->clhash.hashsize; i++) { + hlist_for_each_entry(cl, &q->clhash.hash[i], common.hnode) { if (cl->level) - memset(&cl->un.inner,0,sizeof(cl->un.inner)); + memset(&cl->un.inner, 0, sizeof(cl->un.inner)); else { - if (cl->un.leaf.q) + if (cl->un.leaf.q) qdisc_reset(cl->un.leaf.q); INIT_LIST_HEAD(&cl->un.leaf.drop_list); } cl->prio_activity = 0; cl->cmode = HTB_CAN_SEND; -#ifdef HTB_DEBUG - cl->pq_node.rb_color = -1; - memset(cl->node,255,sizeof(cl->node)); -#endif } } - sch->flags &= ~TCQ_F_THROTTLED; - del_timer(&q->timer); + qdisc_watchdog_cancel(&q->watchdog); __skb_queue_purge(&q->direct_queue); sch->q.qlen = 0; - memset(q->row,0,sizeof(q->row)); - memset(q->row_mask,0,sizeof(q->row_mask)); - memset(q->wait_pq,0,sizeof(q->wait_pq)); - memset(q->ptr,0,sizeof(q->ptr)); + memset(q->hlevel, 0, sizeof(q->hlevel)); + memset(q->row_mask, 0, sizeof(q->row_mask)); for (i = 0; i < TC_HTB_NUMPRIO; i++) - INIT_LIST_HEAD(q->drops+i); + INIT_LIST_HEAD(q->drops + i); +} + +static const struct nla_policy htb_policy[TCA_HTB_MAX + 1] = { + [TCA_HTB_PARMS] = { .len = sizeof(struct tc_htb_opt) }, + [TCA_HTB_INIT] = { .len = sizeof(struct tc_htb_glob) }, + [TCA_HTB_CTAB] = { .type = NLA_BINARY, .len = TC_RTAB_SIZE }, + [TCA_HTB_RTAB] = { .type = NLA_BINARY, .len = TC_RTAB_SIZE }, + [TCA_HTB_DIRECT_QLEN] = { .type = NLA_U32 }, + [TCA_HTB_RATE64] = { .type = NLA_U64 }, + [TCA_HTB_CEIL64] = { .type = NLA_U64 }, +}; + +static void htb_work_func(struct work_struct *work) +{ + struct htb_sched *q = container_of(work, struct htb_sched, work); + struct Qdisc *sch = q->watchdog.qdisc; + + __netif_schedule(qdisc_root(sch)); } -static int htb_init(struct Qdisc *sch, struct rtattr *opt) +static int htb_init(struct Qdisc *sch, struct nlattr *opt) { struct htb_sched *q = qdisc_priv(sch); - struct rtattr *tb[TCA_HTB_INIT]; + struct nlattr *tb[TCA_HTB_MAX + 1]; struct tc_htb_glob *gopt; + int err; int i; -#ifdef HTB_DEBUG - printk(KERN_INFO "HTB init, kernel part version %d.%d\n", - HTB_VER >> 16,HTB_VER & 0xffff); -#endif - if (!opt || rtattr_parse_nested(tb, TCA_HTB_INIT, opt) || - tb[TCA_HTB_INIT-1] == NULL || - RTA_PAYLOAD(tb[TCA_HTB_INIT-1]) < sizeof(*gopt)) { - printk(KERN_ERR "HTB: hey probably you have bad tc tool ?\n"); + + if (!opt) return -EINVAL; - } - gopt = RTA_DATA(tb[TCA_HTB_INIT-1]); - if (gopt->version != HTB_VER >> 16) { - printk(KERN_ERR "HTB: need tc/htb version %d (minor is %d), you have %d\n", - HTB_VER >> 16,HTB_VER & 0xffff,gopt->version); + + err = nla_parse_nested(tb, TCA_HTB_MAX, opt, htb_policy); + if (err < 0) + return err; + + if (!tb[TCA_HTB_INIT]) + return -EINVAL; + + gopt = nla_data(tb[TCA_HTB_INIT]); + if (gopt->version != HTB_VER >> 16) return -EINVAL; - } - q->debug = gopt->debug; - HTB_DBG(0,1,"htb_init sch=%p handle=%X r2q=%d\n",sch,sch->handle,gopt->rate2quantum); - INIT_LIST_HEAD(&q->root); - for (i = 0; i < HTB_HSIZE; i++) - INIT_LIST_HEAD(q->hash+i); + err = qdisc_class_hash_init(&q->clhash); + if (err < 0) + return err; for (i = 0; i < TC_HTB_NUMPRIO; i++) - INIT_LIST_HEAD(q->drops+i); + INIT_LIST_HEAD(q->drops + i); - init_timer(&q->timer); + qdisc_watchdog_init(&q->watchdog, sch); + INIT_WORK(&q->work, htb_work_func); skb_queue_head_init(&q->direct_queue); - q->direct_qlen = sch->dev->tx_queue_len; - if (q->direct_qlen < 2) /* some devices have zero tx_queue_len */ - q->direct_qlen = 2; - q->timer.function = htb_timer; - q->timer.data = (unsigned long)sch; - -#ifdef HTB_RATECM - init_timer(&q->rttim); - q->rttim.function = htb_rate_timer; - q->rttim.data = (unsigned long)sch; - q->rttim.expires = jiffies + HZ; - add_timer(&q->rttim); -#endif + if (tb[TCA_HTB_DIRECT_QLEN]) + q->direct_qlen = nla_get_u32(tb[TCA_HTB_DIRECT_QLEN]); + else { + q->direct_qlen = qdisc_dev(sch)->tx_queue_len; + if (q->direct_qlen < 2) /* some devices have zero tx_queue_len */ + q->direct_qlen = 2; + } if ((q->rate2quantum = gopt->rate2quantum) < 1) q->rate2quantum = 1; q->defcls = gopt->defcls; @@ -1290,89 +1063,89 @@ static int htb_init(struct Qdisc *sch, struct rtattr *opt) static int htb_dump(struct Qdisc *sch, struct sk_buff *skb) { struct htb_sched *q = qdisc_priv(sch); - unsigned char *b = skb->tail; - struct rtattr *rta; + struct nlattr *nest; struct tc_htb_glob gopt; - HTB_DBG(0,1,"htb_dump sch=%p, handle=%X\n",sch,sch->handle); - HTB_QLOCK(sch); - gopt.direct_pkts = q->direct_pkts; -#ifdef HTB_DEBUG - if (HTB_DBG_COND(0,2)) - htb_debug_dump(q); -#endif + /* Its safe to not acquire qdisc lock. As we hold RTNL, + * no change can happen on the qdisc parameters. + */ + + gopt.direct_pkts = q->direct_pkts; gopt.version = HTB_VER; gopt.rate2quantum = q->rate2quantum; gopt.defcls = q->defcls; - gopt.debug = q->debug; - rta = (struct rtattr*)b; - RTA_PUT(skb, TCA_OPTIONS, 0, NULL); - RTA_PUT(skb, TCA_HTB_INIT, sizeof(gopt), &gopt); - rta->rta_len = skb->tail - b; - HTB_QUNLOCK(sch); - return skb->len; -rtattr_failure: - HTB_QUNLOCK(sch); - skb_trim(skb, skb->tail - skb->data); + gopt.debug = 0; + + nest = nla_nest_start(skb, TCA_OPTIONS); + if (nest == NULL) + goto nla_put_failure; + if (nla_put(skb, TCA_HTB_INIT, sizeof(gopt), &gopt) || + nla_put_u32(skb, TCA_HTB_DIRECT_QLEN, q->direct_qlen)) + goto nla_put_failure; + + return nla_nest_end(skb, nest); + +nla_put_failure: + nla_nest_cancel(skb, nest); return -1; } static int htb_dump_class(struct Qdisc *sch, unsigned long arg, - struct sk_buff *skb, struct tcmsg *tcm) + struct sk_buff *skb, struct tcmsg *tcm) { -#ifdef HTB_DEBUG - struct htb_sched *q = qdisc_priv(sch); -#endif - struct htb_class *cl = (struct htb_class*)arg; - unsigned char *b = skb->tail; - struct rtattr *rta; + struct htb_class *cl = (struct htb_class *)arg; + struct nlattr *nest; struct tc_htb_opt opt; - HTB_DBG(0,1,"htb_dump_class handle=%X clid=%X\n",sch->handle,cl->classid); - - HTB_QLOCK(sch); - tcm->tcm_parent = cl->parent ? cl->parent->classid : TC_H_ROOT; - tcm->tcm_handle = cl->classid; + /* Its safe to not acquire qdisc lock. As we hold RTNL, + * no change can happen on the class parameters. + */ + tcm->tcm_parent = cl->parent ? cl->parent->common.classid : TC_H_ROOT; + tcm->tcm_handle = cl->common.classid; if (!cl->level && cl->un.leaf.q) tcm->tcm_info = cl->un.leaf.q->handle; - rta = (struct rtattr*)b; - RTA_PUT(skb, TCA_OPTIONS, 0, NULL); - - memset (&opt,0,sizeof(opt)); - - opt.rate = cl->rate->rate; opt.buffer = cl->buffer; - opt.ceil = cl->ceil->rate; opt.cbuffer = cl->cbuffer; - opt.quantum = cl->un.leaf.quantum; opt.prio = cl->un.leaf.prio; - opt.level = cl->level; - RTA_PUT(skb, TCA_HTB_PARMS, sizeof(opt), &opt); - rta->rta_len = skb->tail - b; - HTB_QUNLOCK(sch); - return skb->len; -rtattr_failure: - HTB_QUNLOCK(sch); - skb_trim(skb, b - skb->data); + nest = nla_nest_start(skb, TCA_OPTIONS); + if (nest == NULL) + goto nla_put_failure; + + memset(&opt, 0, sizeof(opt)); + + psched_ratecfg_getrate(&opt.rate, &cl->rate); + opt.buffer = PSCHED_NS2TICKS(cl->buffer); + psched_ratecfg_getrate(&opt.ceil, &cl->ceil); + opt.cbuffer = PSCHED_NS2TICKS(cl->cbuffer); + opt.quantum = cl->quantum; + opt.prio = cl->prio; + opt.level = cl->level; + if (nla_put(skb, TCA_HTB_PARMS, sizeof(opt), &opt)) + goto nla_put_failure; + if ((cl->rate.rate_bytes_ps >= (1ULL << 32)) && + nla_put_u64(skb, TCA_HTB_RATE64, cl->rate.rate_bytes_ps)) + goto nla_put_failure; + if ((cl->ceil.rate_bytes_ps >= (1ULL << 32)) && + nla_put_u64(skb, TCA_HTB_CEIL64, cl->ceil.rate_bytes_ps)) + goto nla_put_failure; + + return nla_nest_end(skb, nest); + +nla_put_failure: + nla_nest_cancel(skb, nest); return -1; } static int -htb_dump_class_stats(struct Qdisc *sch, unsigned long arg, - struct gnet_dump *d) +htb_dump_class_stats(struct Qdisc *sch, unsigned long arg, struct gnet_dump *d) { - struct htb_class *cl = (struct htb_class*)arg; - -#ifdef HTB_RATECM - cl->rate_est.bps = cl->rate_bytes/(HTB_EWMAC*HTB_HSIZE); - cl->rate_est.pps = cl->rate_packets/(HTB_EWMAC*HTB_HSIZE); -#endif + struct htb_class *cl = (struct htb_class *)arg; if (!cl->level && cl->un.leaf.q) cl->qstats.qlen = cl->un.leaf.q->q.qlen; - cl->xstats.tokens = cl->tokens; - cl->xstats.ctokens = cl->ctokens; + cl->xstats.tokens = PSCHED_NS2TICKS(cl->tokens); + cl->xstats.ctokens = PSCHED_NS2TICKS(cl->ctokens); if (gnet_stats_copy_basic(d, &cl->bstats) < 0 || - gnet_stats_copy_rate_est(d, &cl->rate_est) < 0 || + gnet_stats_copy_rate_est(d, NULL, &cl->rate_est) < 0 || gnet_stats_copy_queue(d, &cl->qstats) < 0) return -1; @@ -1380,132 +1153,171 @@ htb_dump_class_stats(struct Qdisc *sch, unsigned long arg, } static int htb_graft(struct Qdisc *sch, unsigned long arg, struct Qdisc *new, - struct Qdisc **old) + struct Qdisc **old) { - struct htb_class *cl = (struct htb_class*)arg; + struct htb_class *cl = (struct htb_class *)arg; - if (cl && !cl->level) { - if (new == NULL && (new = qdisc_create_dflt(sch->dev, - &pfifo_qdisc_ops)) == NULL) - return -ENOBUFS; - sch_tree_lock(sch); - if ((*old = xchg(&cl->un.leaf.q, new)) != NULL) { - if (cl->prio_activity) - htb_deactivate (qdisc_priv(sch),cl); + if (cl->level) + return -EINVAL; + if (new == NULL && + (new = qdisc_create_dflt(sch->dev_queue, &pfifo_qdisc_ops, + cl->common.classid)) == NULL) + return -ENOBUFS; - /* TODO: is it correct ? Why CBQ doesn't do it ? */ - sch->q.qlen -= (*old)->q.qlen; - qdisc_reset(*old); - } - sch_tree_unlock(sch); - return 0; + sch_tree_lock(sch); + *old = cl->un.leaf.q; + cl->un.leaf.q = new; + if (*old != NULL) { + qdisc_tree_decrease_qlen(*old, (*old)->q.qlen); + qdisc_reset(*old); } - return -ENOENT; + sch_tree_unlock(sch); + return 0; +} + +static struct Qdisc *htb_leaf(struct Qdisc *sch, unsigned long arg) +{ + struct htb_class *cl = (struct htb_class *)arg; + return !cl->level ? cl->un.leaf.q : NULL; } -static struct Qdisc * htb_leaf(struct Qdisc *sch, unsigned long arg) +static void htb_qlen_notify(struct Qdisc *sch, unsigned long arg) { - struct htb_class *cl = (struct htb_class*)arg; - return (cl && !cl->level) ? cl->un.leaf.q : NULL; + struct htb_class *cl = (struct htb_class *)arg; + + if (cl->un.leaf.q->q.qlen == 0) + htb_deactivate(qdisc_priv(sch), cl); } static unsigned long htb_get(struct Qdisc *sch, u32 classid) { -#ifdef HTB_DEBUG - struct htb_sched *q = qdisc_priv(sch); -#endif - struct htb_class *cl = htb_find(classid,sch); - HTB_DBG(0,1,"htb_get clid=%X q=%p cl=%p ref=%d\n",classid,q,cl,cl?cl->refcnt:0); - if (cl) + struct htb_class *cl = htb_find(classid, sch); + if (cl) cl->refcnt++; return (unsigned long)cl; } -static void htb_destroy_filters(struct tcf_proto **fl) +static inline int htb_parent_last_child(struct htb_class *cl) { - struct tcf_proto *tp; + if (!cl->parent) + /* the root class */ + return 0; + if (cl->parent->children > 1) + /* not the last child */ + return 0; + return 1; +} - while ((tp = *fl) != NULL) { - *fl = tp->next; - tcf_destroy(tp); - } +static void htb_parent_to_leaf(struct htb_sched *q, struct htb_class *cl, + struct Qdisc *new_q) +{ + struct htb_class *parent = cl->parent; + + WARN_ON(cl->level || !cl->un.leaf.q || cl->prio_activity); + + if (parent->cmode != HTB_CAN_SEND) + htb_safe_rb_erase(&parent->pq_node, + &q->hlevel[parent->level].wait_pq); + + parent->level = 0; + memset(&parent->un.inner, 0, sizeof(parent->un.inner)); + INIT_LIST_HEAD(&parent->un.leaf.drop_list); + parent->un.leaf.q = new_q ? new_q : &noop_qdisc; + parent->tokens = parent->buffer; + parent->ctokens = parent->cbuffer; + parent->t_c = ktime_to_ns(ktime_get()); + parent->cmode = HTB_CAN_SEND; } -static void htb_destroy_class(struct Qdisc* sch,struct htb_class *cl) +static void htb_destroy_class(struct Qdisc *sch, struct htb_class *cl) { - struct htb_sched *q = qdisc_priv(sch); - HTB_DBG(0,1,"htb_destrycls clid=%X ref=%d\n", cl?cl->classid:0,cl?cl->refcnt:0); if (!cl->level) { - BUG_TRAP(cl->un.leaf.q); - sch->q.qlen -= cl->un.leaf.q->q.qlen; + WARN_ON(!cl->un.leaf.q); qdisc_destroy(cl->un.leaf.q); } - qdisc_put_rtab(cl->rate); - qdisc_put_rtab(cl->ceil); - - htb_destroy_filters (&cl->filter_list); - - while (!list_empty(&cl->children)) - htb_destroy_class (sch,list_entry(cl->children.next, - struct htb_class,sibling)); - - /* note: this delete may happen twice (see htb_delete) */ - list_del(&cl->hlist); - list_del(&cl->sibling); - - if (cl->prio_activity) - htb_deactivate (q,cl); - - if (cl->cmode != HTB_CAN_SEND) - htb_safe_rb_erase(&cl->pq_node,q->wait_pq+cl->level); - + gen_kill_estimator(&cl->bstats, &cl->rate_est); + tcf_destroy_chain(&cl->filter_list); kfree(cl); } -/* always caled under BH & queue lock */ -static void htb_destroy(struct Qdisc* sch) +static void htb_destroy(struct Qdisc *sch) { struct htb_sched *q = qdisc_priv(sch); - HTB_DBG(0,1,"htb_destroy q=%p\n",q); + struct hlist_node *next; + struct htb_class *cl; + unsigned int i; - del_timer_sync (&q->timer); -#ifdef HTB_RATECM - del_timer_sync (&q->rttim); -#endif + cancel_work_sync(&q->work); + qdisc_watchdog_cancel(&q->watchdog); /* This line used to be after htb_destroy_class call below - and surprisingly it worked in 2.4. But it must precede it - because filter need its target class alive to be able to call - unbind_filter on it (without Oops). */ - htb_destroy_filters(&q->filter_list); - - while (!list_empty(&q->root)) - htb_destroy_class (sch,list_entry(q->root.next, - struct htb_class,sibling)); + * and surprisingly it worked in 2.4. But it must precede it + * because filter need its target class alive to be able to call + * unbind_filter on it (without Oops). + */ + tcf_destroy_chain(&q->filter_list); + for (i = 0; i < q->clhash.hashsize; i++) { + hlist_for_each_entry(cl, &q->clhash.hash[i], common.hnode) + tcf_destroy_chain(&cl->filter_list); + } + for (i = 0; i < q->clhash.hashsize; i++) { + hlist_for_each_entry_safe(cl, next, &q->clhash.hash[i], + common.hnode) + htb_destroy_class(sch, cl); + } + qdisc_class_hash_destroy(&q->clhash); __skb_queue_purge(&q->direct_queue); } static int htb_delete(struct Qdisc *sch, unsigned long arg) { struct htb_sched *q = qdisc_priv(sch); - struct htb_class *cl = (struct htb_class*)arg; - HTB_DBG(0,1,"htb_delete q=%p cl=%X ref=%d\n",q,cl?cl->classid:0,cl?cl->refcnt:0); + struct htb_class *cl = (struct htb_class *)arg; + unsigned int qlen; + struct Qdisc *new_q = NULL; + int last_child = 0; - // TODO: why don't allow to delete subtree ? references ? does - // tc subsys quarantee us that in htb_destroy it holds no class - // refs so that we can remove children safely there ? - if (!list_empty(&cl->children) || cl->filter_cnt) + /* TODO: why don't allow to delete subtree ? references ? does + * tc subsys guarantee us that in htb_destroy it holds no class + * refs so that we can remove children safely there ? + */ + if (cl->children || cl->filter_cnt) return -EBUSY; - + + if (!cl->level && htb_parent_last_child(cl)) { + new_q = qdisc_create_dflt(sch->dev_queue, &pfifo_qdisc_ops, + cl->parent->common.classid); + last_child = 1; + } + sch_tree_lock(sch); - + + if (!cl->level) { + qlen = cl->un.leaf.q->q.qlen; + qdisc_reset(cl->un.leaf.q); + qdisc_tree_decrease_qlen(cl->un.leaf.q, qlen); + } + /* delete from hash and active; remainder in destroy_class */ - list_del_init(&cl->hlist); + qdisc_class_hash_remove(&q->clhash, &cl->common); + if (cl->parent) + cl->parent->children--; + if (cl->prio_activity) - htb_deactivate (q,cl); + htb_deactivate(q, cl); - if (--cl->refcnt == 0) - htb_destroy_class(sch,cl); + if (cl->cmode != HTB_CAN_SEND) + htb_safe_rb_erase(&cl->pq_node, + &q->hlevel[cl->level].wait_pq); + + if (last_child) + htb_parent_to_leaf(q, cl, new_q); + + BUG_ON(--cl->refcnt == 0); + /* + * This shouldn't happen: we "hold" one cops->get() when called + * from tc_ctl_tclass; the destroy method is done from cops->put(). + */ sch_tree_unlock(sch); return 0; @@ -1513,141 +1325,197 @@ static int htb_delete(struct Qdisc *sch, unsigned long arg) static void htb_put(struct Qdisc *sch, unsigned long arg) { -#ifdef HTB_DEBUG - struct htb_sched *q = qdisc_priv(sch); -#endif - struct htb_class *cl = (struct htb_class*)arg; - HTB_DBG(0,1,"htb_put q=%p cl=%X ref=%d\n",q,cl?cl->classid:0,cl?cl->refcnt:0); + struct htb_class *cl = (struct htb_class *)arg; if (--cl->refcnt == 0) - htb_destroy_class(sch,cl); + htb_destroy_class(sch, cl); } -static int htb_change_class(struct Qdisc *sch, u32 classid, - u32 parentid, struct rtattr **tca, unsigned long *arg) +static int htb_change_class(struct Qdisc *sch, u32 classid, + u32 parentid, struct nlattr **tca, + unsigned long *arg) { int err = -EINVAL; struct htb_sched *q = qdisc_priv(sch); - struct htb_class *cl = (struct htb_class*)*arg,*parent; - struct rtattr *opt = tca[TCA_OPTIONS-1]; - struct qdisc_rate_table *rtab = NULL, *ctab = NULL; - struct rtattr *tb[TCA_HTB_RTAB]; + struct htb_class *cl = (struct htb_class *)*arg, *parent; + struct nlattr *opt = tca[TCA_OPTIONS]; + struct nlattr *tb[TCA_HTB_MAX + 1]; struct tc_htb_opt *hopt; + u64 rate64, ceil64; /* extract all subattrs from opt attr */ - if (!opt || rtattr_parse_nested(tb, TCA_HTB_RTAB, opt) || - tb[TCA_HTB_PARMS-1] == NULL || - RTA_PAYLOAD(tb[TCA_HTB_PARMS-1]) < sizeof(*hopt)) + if (!opt) goto failure; - - parent = parentid == TC_H_ROOT ? NULL : htb_find (parentid,sch); - hopt = RTA_DATA(tb[TCA_HTB_PARMS-1]); - HTB_DBG(0,1,"htb_chg cl=%p(%X), clid=%X, parid=%X, opt/prio=%d, rate=%u, buff=%d, quant=%d\n", cl,cl?cl->classid:0,classid,parentid,(int)hopt->prio,hopt->rate.rate,hopt->buffer,hopt->quantum); - rtab = qdisc_get_rtab(&hopt->rate, tb[TCA_HTB_RTAB-1]); - ctab = qdisc_get_rtab(&hopt->ceil, tb[TCA_HTB_CTAB-1]); - if (!rtab || !ctab) goto failure; + err = nla_parse_nested(tb, TCA_HTB_MAX, opt, htb_policy); + if (err < 0) + goto failure; - if (!cl) { /* new class */ + err = -EINVAL; + if (tb[TCA_HTB_PARMS] == NULL) + goto failure; + + parent = parentid == TC_H_ROOT ? NULL : htb_find(parentid, sch); + + hopt = nla_data(tb[TCA_HTB_PARMS]); + if (!hopt->rate.rate || !hopt->ceil.rate) + goto failure; + + /* Keeping backward compatible with rate_table based iproute2 tc */ + if (hopt->rate.linklayer == TC_LINKLAYER_UNAWARE) + qdisc_put_rtab(qdisc_get_rtab(&hopt->rate, tb[TCA_HTB_RTAB])); + + if (hopt->ceil.linklayer == TC_LINKLAYER_UNAWARE) + qdisc_put_rtab(qdisc_get_rtab(&hopt->ceil, tb[TCA_HTB_CTAB])); + + if (!cl) { /* new class */ struct Qdisc *new_q; + int prio; + struct { + struct nlattr nla; + struct gnet_estimator opt; + } est = { + .nla = { + .nla_len = nla_attr_size(sizeof(est.opt)), + .nla_type = TCA_RATE, + }, + .opt = { + /* 4s interval, 16s averaging constant */ + .interval = 2, + .ewma_log = 2, + }, + }; + /* check for valid classid */ - if (!classid || TC_H_MAJ(classid^sch->handle) || htb_find(classid,sch)) + if (!classid || TC_H_MAJ(classid ^ sch->handle) || + htb_find(classid, sch)) goto failure; /* check maximal depth */ if (parent && parent->parent && parent->parent->level < 2) { - printk(KERN_ERR "htb: tree is too deep\n"); + pr_err("htb: tree is too deep\n"); goto failure; } err = -ENOBUFS; - if ((cl = kmalloc(sizeof(*cl), GFP_KERNEL)) == NULL) + cl = kzalloc(sizeof(*cl), GFP_KERNEL); + if (!cl) goto failure; - - memset(cl, 0, sizeof(*cl)); + + if (htb_rate_est || tca[TCA_RATE]) { + err = gen_new_estimator(&cl->bstats, &cl->rate_est, + qdisc_root_sleeping_lock(sch), + tca[TCA_RATE] ? : &est.nla); + if (err) { + kfree(cl); + goto failure; + } + } + cl->refcnt = 1; - INIT_LIST_HEAD(&cl->sibling); - INIT_LIST_HEAD(&cl->hlist); - INIT_LIST_HEAD(&cl->children); + cl->children = 0; INIT_LIST_HEAD(&cl->un.leaf.drop_list); -#ifdef HTB_DEBUG - cl->magic = HTB_CMAGIC; -#endif + RB_CLEAR_NODE(&cl->pq_node); + + for (prio = 0; prio < TC_HTB_NUMPRIO; prio++) + RB_CLEAR_NODE(&cl->node[prio]); /* create leaf qdisc early because it uses kmalloc(GFP_KERNEL) - so that can't be used inside of sch_tree_lock - -- thanks to Karlis Peisenieks */ - new_q = qdisc_create_dflt(sch->dev, &pfifo_qdisc_ops); + * so that can't be used inside of sch_tree_lock + * -- thanks to Karlis Peisenieks + */ + new_q = qdisc_create_dflt(sch->dev_queue, + &pfifo_qdisc_ops, classid); sch_tree_lock(sch); if (parent && !parent->level) { + unsigned int qlen = parent->un.leaf.q->q.qlen; + /* turn parent into inner node */ - sch->q.qlen -= parent->un.leaf.q->q.qlen; - qdisc_destroy (parent->un.leaf.q); - if (parent->prio_activity) - htb_deactivate (q,parent); + qdisc_reset(parent->un.leaf.q); + qdisc_tree_decrease_qlen(parent->un.leaf.q, qlen); + qdisc_destroy(parent->un.leaf.q); + if (parent->prio_activity) + htb_deactivate(q, parent); /* remove from evt list because of level change */ if (parent->cmode != HTB_CAN_SEND) { - htb_safe_rb_erase(&parent->pq_node,q->wait_pq /*+0*/); + htb_safe_rb_erase(&parent->pq_node, &q->hlevel[0].wait_pq); parent->cmode = HTB_CAN_SEND; } parent->level = (parent->parent ? parent->parent->level - : TC_HTB_MAXDEPTH) - 1; - memset (&parent->un.inner,0,sizeof(parent->un.inner)); + : TC_HTB_MAXDEPTH) - 1; + memset(&parent->un.inner, 0, sizeof(parent->un.inner)); } /* leaf (we) needs elementary qdisc */ cl->un.leaf.q = new_q ? new_q : &noop_qdisc; - cl->classid = classid; cl->parent = parent; + cl->common.classid = classid; + cl->parent = parent; /* set class to be in HTB_CAN_SEND state */ - cl->tokens = hopt->buffer; - cl->ctokens = hopt->cbuffer; - cl->mbuffer = 60000000; /* 1min */ - PSCHED_GET_TIME(cl->t_c); + cl->tokens = PSCHED_TICKS2NS(hopt->buffer); + cl->ctokens = PSCHED_TICKS2NS(hopt->cbuffer); + cl->mbuffer = 60ULL * NSEC_PER_SEC; /* 1min */ + cl->t_c = ktime_to_ns(ktime_get()); cl->cmode = HTB_CAN_SEND; /* attach to the hash list and parent's family */ - list_add_tail(&cl->hlist, q->hash+htb_hash(classid)); - list_add_tail(&cl->sibling, parent ? &parent->children : &q->root); -#ifdef HTB_DEBUG - { - int i; - for (i = 0; i < TC_HTB_NUMPRIO; i++) cl->node[i].rb_color = -1; - cl->pq_node.rb_color = -1; + qdisc_class_hash_insert(&q->clhash, &cl->common); + if (parent) + parent->children++; + } else { + if (tca[TCA_RATE]) { + err = gen_replace_estimator(&cl->bstats, &cl->rate_est, + qdisc_root_sleeping_lock(sch), + tca[TCA_RATE]); + if (err) + return err; } -#endif - } else sch_tree_lock(sch); + sch_tree_lock(sch); + } + + rate64 = tb[TCA_HTB_RATE64] ? nla_get_u64(tb[TCA_HTB_RATE64]) : 0; + + ceil64 = tb[TCA_HTB_CEIL64] ? nla_get_u64(tb[TCA_HTB_CEIL64]) : 0; + + psched_ratecfg_precompute(&cl->rate, &hopt->rate, rate64); + psched_ratecfg_precompute(&cl->ceil, &hopt->ceil, ceil64); /* it used to be a nasty bug here, we have to check that node - is really leaf before changing cl->un.leaf ! */ + * is really leaf before changing cl->un.leaf ! + */ if (!cl->level) { - cl->un.leaf.quantum = rtab->rate.rate / q->rate2quantum; - if (!hopt->quantum && cl->un.leaf.quantum < 1000) { - printk(KERN_WARNING "HTB: quantum of class %X is small. Consider r2q change.\n", cl->classid); - cl->un.leaf.quantum = 1000; + u64 quantum = cl->rate.rate_bytes_ps; + + do_div(quantum, q->rate2quantum); + cl->quantum = min_t(u64, quantum, INT_MAX); + + if (!hopt->quantum && cl->quantum < 1000) { + pr_warn("HTB: quantum of class %X is small. Consider r2q change.\n", + cl->common.classid); + cl->quantum = 1000; } - if (!hopt->quantum && cl->un.leaf.quantum > 200000) { - printk(KERN_WARNING "HTB: quantum of class %X is big. Consider r2q change.\n", cl->classid); - cl->un.leaf.quantum = 200000; + if (!hopt->quantum && cl->quantum > 200000) { + pr_warn("HTB: quantum of class %X is big. Consider r2q change.\n", + cl->common.classid); + cl->quantum = 200000; } if (hopt->quantum) - cl->un.leaf.quantum = hopt->quantum; - if ((cl->un.leaf.prio = hopt->prio) >= TC_HTB_NUMPRIO) - cl->un.leaf.prio = TC_HTB_NUMPRIO - 1; + cl->quantum = hopt->quantum; + if ((cl->prio = hopt->prio) >= TC_HTB_NUMPRIO) + cl->prio = TC_HTB_NUMPRIO - 1; } - cl->buffer = hopt->buffer; - cl->cbuffer = hopt->cbuffer; - if (cl->rate) qdisc_put_rtab(cl->rate); cl->rate = rtab; - if (cl->ceil) qdisc_put_rtab(cl->ceil); cl->ceil = ctab; + cl->buffer = PSCHED_TICKS2NS(hopt->buffer); + cl->cbuffer = PSCHED_TICKS2NS(hopt->cbuffer); + sch_tree_unlock(sch); + qdisc_class_hash_grow(sch, &q->clhash); + *arg = (unsigned long)cl; return 0; failure: - if (rtab) qdisc_put_rtab(rtab); - if (ctab) qdisc_put_rtab(ctab); return err; } @@ -1656,55 +1524,48 @@ static struct tcf_proto **htb_find_tcf(struct Qdisc *sch, unsigned long arg) struct htb_sched *q = qdisc_priv(sch); struct htb_class *cl = (struct htb_class *)arg; struct tcf_proto **fl = cl ? &cl->filter_list : &q->filter_list; - HTB_DBG(0,2,"htb_tcf q=%p clid=%X fref=%d fl=%p\n",q,cl?cl->classid:0,cl?cl->filter_cnt:q->filter_cnt,*fl); + return fl; } static unsigned long htb_bind_filter(struct Qdisc *sch, unsigned long parent, - u32 classid) + u32 classid) { - struct htb_sched *q = qdisc_priv(sch); - struct htb_class *cl = htb_find (classid,sch); - HTB_DBG(0,2,"htb_bind q=%p clid=%X cl=%p fref=%d\n",q,classid,cl,cl?cl->filter_cnt:q->filter_cnt); + struct htb_class *cl = htb_find(classid, sch); + /*if (cl && !cl->level) return 0; - The line above used to be there to prevent attaching filters to - leaves. But at least tc_index filter uses this just to get class - for other reasons so that we have to allow for it. - ---- - 19.6.2002 As Werner explained it is ok - bind filter is just - another way to "lock" the class - unlike "get" this lock can - be broken by class during destroy IIUC. + * The line above used to be there to prevent attaching filters to + * leaves. But at least tc_index filter uses this just to get class + * for other reasons so that we have to allow for it. + * ---- + * 19.6.2002 As Werner explained it is ok - bind filter is just + * another way to "lock" the class - unlike "get" this lock can + * be broken by class during destroy IIUC. */ - if (cl) - cl->filter_cnt++; - else - q->filter_cnt++; + if (cl) + cl->filter_cnt++; return (unsigned long)cl; } static void htb_unbind_filter(struct Qdisc *sch, unsigned long arg) { - struct htb_sched *q = qdisc_priv(sch); struct htb_class *cl = (struct htb_class *)arg; - HTB_DBG(0,2,"htb_unbind q=%p cl=%p fref=%d\n",q,cl,cl?cl->filter_cnt:q->filter_cnt); - if (cl) - cl->filter_cnt--; - else - q->filter_cnt--; + + if (cl) + cl->filter_cnt--; } static void htb_walk(struct Qdisc *sch, struct qdisc_walker *arg) { struct htb_sched *q = qdisc_priv(sch); - int i; + struct htb_class *cl; + unsigned int i; if (arg->stop) return; - for (i = 0; i < HTB_HSIZE; i++) { - struct list_head *p; - list_for_each (p,q->hash+i) { - struct htb_class *cl = list_entry(p,struct htb_class,hlist); + for (i = 0; i < q->clhash.hashsize; i++) { + hlist_for_each_entry(cl, &q->clhash.hash[i], common.hnode) { if (arg->count < arg->skip) { arg->count++; continue; @@ -1718,9 +1579,10 @@ static void htb_walk(struct Qdisc *sch, struct qdisc_walker *arg) } } -static struct Qdisc_class_ops htb_class_ops = { +static const struct Qdisc_class_ops htb_class_ops = { .graft = htb_graft, .leaf = htb_leaf, + .qlen_notify = htb_qlen_notify, .get = htb_get, .put = htb_put, .change = htb_change_class, @@ -1733,31 +1595,30 @@ static struct Qdisc_class_ops htb_class_ops = { .dump_stats = htb_dump_class_stats, }; -static struct Qdisc_ops htb_qdisc_ops = { - .next = NULL, +static struct Qdisc_ops htb_qdisc_ops __read_mostly = { .cl_ops = &htb_class_ops, .id = "htb", .priv_size = sizeof(struct htb_sched), .enqueue = htb_enqueue, .dequeue = htb_dequeue, - .requeue = htb_requeue, + .peek = qdisc_peek_dequeued, .drop = htb_drop, .init = htb_init, .reset = htb_reset, .destroy = htb_destroy, - .change = NULL /* htb_change */, .dump = htb_dump, .owner = THIS_MODULE, }; static int __init htb_module_init(void) { - return register_qdisc(&htb_qdisc_ops); + return register_qdisc(&htb_qdisc_ops); } -static void __exit htb_module_exit(void) +static void __exit htb_module_exit(void) { - unregister_qdisc(&htb_qdisc_ops); + unregister_qdisc(&htb_qdisc_ops); } + module_init(htb_module_init) module_exit(htb_module_exit) MODULE_LICENSE("GPL"); |